Impaired CD4+ T cell restoration in the small vs. large intestine of HIV-1+ South Africans receiving...

M A J O R A R T I C L E

Impaired CD4+ T-Cell Restoration in the SmallVersus Large Intestine of HIV-1–Positive SouthAfricans Receiving Combination AntiretroviralTherapy

Edana Cassol,1,2,3,b Susan Malfeld,3,7 Phetole Mahasha,3,7 Robert Bond,4 Tomas Slavik,5 Chris Seebregts,8 Guido Poli,9,10

Sharon Cassol,3,7 Schalk W. van der Merwe,3,4,7,a and Theresa Rossouw6,a

1Department of Cancer Immunology and AIDS, Dana-Farber Cancer Institute, and 2Department of Microbiology and Immunobiology, Harvard MedicalSchool, Boston, Massachusetts; 3Medical Research Council Unit for Inflammation and Immunity, Department of Immunology, 4Hepatology andGastrointestinal Research Laboratory, 5Ampath Pathology Laboratories, Pretoria, Department of Anatomical Pathology, and 6Department of FamilyMedicine, University of Pretoria, Pretoria, 7Tshwane Academic Division, National Health Laboratory Service, Johannesburg, and 8Biomedical InformaticsResearch, eHealth Research and Innovation Platform, Medical Research Council, Tygerberg, South Africa; and 9AIDS Immunopathogenesis Unit, Divisionof Immunology, Transplantation, and Infectious Diseases, San Raffaele Scientific Institute, and 10Vita-Salute San Raffaele University, School ofMedicine, Milan, Italy

Background. Human immunodeficiency virus type 1 (HIV-1) infection is associated with a massive depletionof intestinal CD4+ T cells that is only partially reversed by combination antiretroviral therapy (cART). Here, we as-sessed the ability of nucleoside reverse-transcriptase inhibitor/nonnucleoside reverse-transcriptase inhibitor treat-ment to restore the CD4+ T-cell populations in the intestine of South African patients with AIDS.

Methods. Thirty-eight patients with advanced HIV-1 infection who had chronic diarrhea (duration, >4 weeks)and/or unintentional weight loss (>10% decrease from baseline) of uncertain etiology were enrolled. Blood specimenswere collected monthly, and gastrointestinal tract biopsy specimens were collected before cART initiation (from theduodenum, jejunum, ileum, and colon), 3 months after cART initiation (from the duodenum), and 6 months aftercART initiation (from the duodenum and colon). CD4+, CD8+, and CD38+CD8+ T cells were quantified by flow cy-tometry and immunohistochemistry analyses, and the HIV-1 RNA load was determined by the Nuclisens assay.

Results. CD4+ T-cell and HIV-1 RNA levels were significantly lower, whereas CD8+ T-cell levels, including acti-vated CD38+CD8+ T cell levels, were higher in the duodenum and jejunum, compared with the colon. After 6 monthsof cART, a significant but incomplete recovery of CD4+ T cells was detected in the colon and peripheral blood but notin the duodenum. Failed restoration of the CD4+ T-cell count in the duodenum was associated with nonspecific enteri-tis and CD8+ T-cell activation.

Conclusions. Strategies that target inflammation and immune activation in the small intestine may be required toexpedite CD4+ T-cell recovery and improve therapeutic outcomes.

Keywords. HIV-1; antiretroviral therapy; CD4 reconstitution; intestine; Africa; immune activation.

Human immunodeficiency virus type 1 (HIV-1) infec-tion is characterized by a profound depletion of CCR5+

memory CD4+ T cells in the lamina propria of the gas-trointestinal tract [1–3]. This loss has been attributed tothe direct killing of these cells by HIV-1 and to bystand-er apoptosis of uninfected cells [4, 5]. Depletion of CD4+

T cells in the gastrointestinal tract has been implicated inthe immune dysfunction and loss of epithelial integritythat leads to increased microbial translocation, immuneactivation, and disease progression [6–8].

Received 11 September 2012; accepted 27 December 2012; electronically pub-lished 6 June 2013.

Presented in part: 14th International Congress of Mucosal Immunology, 5–9 July2009, Boston, Massachusetts.

aT. R. and S. W. v. d. M. contributed equally to this work.bPresent affiliation: Department of Cancer Immunology and AIDS, Dana-Farber

Cancer Institute, Boston, Massachusetts.Correspondence: Edana Cassol, PhD, Department of Cancer Immunology and

AIDS, Dana-Farber Cancer Institute, 450 Brookline Ave, CLS1017 Boston MA 02215([email protected]).

The Journal of Infectious Diseases 2013;208:1113–22© The Author 2013. Published by Oxford University Press on behalf of the InfectiousDiseases Society of America. All rights reserved. For Permissions, please e-mail:[email protected]: 10.1093/infdis/jit249

cART-Induced Recovery of Gut CD4+ T Cells • JID 2013:208 (1 October) • 1113

at Harvard U

niversity on September 27, 2013

http://jid.oxfordjournals.org/D

ownloaded from

at H

arvard University on Septem

ber 27, 2013http://jid.oxfordjournals.org/

Dow

nloaded from

at Harvard U



ownloaded from

at H



Dow

nloaded from

at Harvard U



ownloaded from

at H



Dow

nloaded from

at Harvard U



ownloaded from

at H



Dow

nloaded from

at Harvard U



ownloaded from

at H



Dow

nloaded from

mailto:[email protected]

mailto:[email protected]

http://jid.oxfordjournals.org/










There are still many unanswered questions relating to theloss of intestinal CD4+ T cells and HIV-1 pathogenesis, partic-ularly since intestinal CD4+ T cells are also depleted in non-pathogenic simian immunodeficiency virus (SIV) infection [9, 10].Recent studies suggest differential depletion of T-helper 17 CD4+

T cells may account for this discrepancy as these cells are pref-erentially depleted in HIV-1 and pathogenic SIV infection andpreserved in nonpathogenic SIV infection [11]. Importantly, T-helper 17 cells play a key role in regulating intestinal immunityand in maintaining the mucosal epithelium [12, 13]. Strategiesaimed at restoring the CD4+ T-cell count in the gastrointestinaltract may be critical for reestablishing epithelial integrity andreducing chronic inflammation and immune activation.

Findings of investigations to determine whether combinationantiretroviral therapy (cART) can restore the CD4+ T-cell levelin the gastrointestinal tract have, thus far, been inconclusive.Most studies have been cross-sectional, have been performed ata single site, and have involved a small number of patients.Overall, these studies suggest that restoration of the CD4+ T-cellcount in the gastrointestinal tract is incomplete and occurs moreslowly than in blood, with many patients exhibiting limitedrestoration despite prolonged suppressive treatment [14–17].Other studies have reported near-complete CD4+ T-cell countreconstitution in the jejunum of patients treated during acuteinfection [18] and in the ileum [19] and colon [19, 20] of pa-tients receiving long-term (duration, >4 years) suppressivecART. Further investigation is needed to determine whetherthese discrepancies are due to differences in the timing, duration,and type of treatment or to tissue-specific differences in homingreceptors, immune activation, and/or collagen deposition.

To assess how cART affects immune reconstitution in differ-ent regions of the gastrointestinal tract, we compared CD4+,CD8+, and CD38+CD8+ T-cell levels and HIV-1 RNA levels inintestinal biopsy specimens from South African patients withlate-stage infection who had chronic diarrhea and/or weightloss. Restoration of the CD4+ T-cell levels during cART was im-paired in the upper small intestine (ie, the duodenum andjejunum), compared with the colon and blood. Failed restora-tion of the CD4+ T-cell count was associated with nonspecificenteritis and sustained CD8+ T-cell activation. Treatment strat-egies that target inflammation and immune activation in theupper gastrointestinal tract may provide significant clinical andimmunological benefit.

MATERIAL ANDMETHODS

Study ParticipantsThirty-eight patients with advanced HIV-1 infection who hadchronic diarrhea (duration, >4 weeks) and/or unintentionalweight loss (>10% decrease from baseline) of uncertain eti-ology were recruited at the Comprehensive Care, Manage-ment, and Treatment Clinic in Pretoria, South Africa. Following

gastrointestinal biopsy, they immediately initiated a standard-ized first-line drug regimen consisting of 2 nucleoside reverse-transcriptase inhibitors (NRTIs; stavudine plus lamivudine) and1 nonnucleoside reverse-transcriptase inhibitor (NNRTI; eitherefavirenz or nevirapine), as recommended by South Africanguidelines [21]. Biopsy specimens were collected at enrollment(from the duodenum, jejunum, ileum, and colon), after 3months of cART (from the duodenum), and after 6 months ofcART (from the duodenum and colon). Monthly blood sampleswere collected for measurement of CD4+ and CD8+ T cellcounts and viral load. Patients were eligible for the study if acid-fast stains of sputum smears and stool specimens were negativefor Mycobacterium tuberculosis and enteric pathogens, respec-tively. Control biopsy specimens were obtained from the duode-num and colon of 5 HIV-1–seronegative controls who testednegative for pathogens and showed no evidence of disease duringscreening for colorectal cancer. Written informed consent wasobtained from all study participants. The study was approved bythe Ethics Committee of the University of Pretoria.

Sample CollectionBiopsy specimens from the duodenum and jejunum were col-lected using a Fujinon double-balloon enteroscope and RadialIII biopsy forceps (Boston Scientific). Endoscopic biopsy speci-mens from the colon and terminal ileum were obtainedthrough the anal route. A total of 8–10 samples were randomlycollected from each site. Biopsy specimens were placed on icein Roswell Park Memorial Institute (RPMI) 1640 medium with10% fetal calf serum and immediately processed for flow cytom-etry, fixed in 10% formalin for immunohistochemistry analysis,or snap frozen for tissue-associated HIV-1 RNA load analysis.

ImmunophenotypingMultiparameter flow cytometry was performed on mononucle-ar cells isolated from intestinal biopsy specimens. Tissuesamples were digested (for 30 minutes at 37°C) with collage-nase type IV (0.5 mg/mL; Sigma, St. Louis, MO) in RPMI 1640medium and passed through a 70-μm cell strainer (BectonDickinson Labware, Franklin Lakes, NJ). Residual tissue frag-ments were redigested, and the pooled single cell suspensionswere washed to remove collagenase. Mononuclear cells werestained (for 30 minutes at 4°C) with various combinations ofanti-human CD4, CD8, CD3, CD45, CD45RO, CD27, andCD38 monoclonal antibodies (mAb; Beckman Coulter). Thepercentage of CD4+, CD8+, CD38+CD8+, CD45RO+CD4+

memory, and CD45RO−CD27+CD4+ naive T cells was deter-mined by gating on the lymphocyte population defined byforward- and side-scatter characteristics (confirmed to beCD45+CD3+ cells by back gating), followed by sequentialgating for either CD4 or CD8. A minimum of 5000 lymphocyteevents were collected per tube on a Beckman Coulter FC500.Results were analyzed using Flow Jo (Tree Star, Ashland, OR).

1114 • JID 2013:208 (1 October) • Cassol et al

Histological AnalysisStaining with hematoxylin-eosin, periodic acid-Schiff diastase,Ziehl-Neelsen, and Giemsa was performed to exclude fungaland parasitic organisms (Cryptosporidium, Isospora, and Mi-crosporidia) and acid-fast bacilli. For T-cell subset quantifica-tion, sections of the duodenum and colon were incubated withmurine anti-human CD4 or CD8 mAbs (Dako, Denmark), vi-sualized with diaminobenzidine, and counterstained with he-matoxylin. The number of CD4+ and CD8+ lymphocytes in 5representative high-power fields of 0.80 mm2 (original magnifi-cation, ×400) was counted, and the results are reported as theaverage number of cells/0.8 mm2. All histological studies wereperformed in a blinded fashion by a histopathologist with exten-sive experience in the analysis of intestinal biopsy specimens.

Tissue-Associated Viral LoadBiopsy specimens (4–12 mg) were weighed, minced with arazor blade, and placed in lysis buffer (BioMerieux) overnightat room temperature. RNA was extracted using the MagneticExtraction Reagent Kit from BioMerieux. Viral RNA in tissueextracts and plasma was quantified using the Nuclisens Easy QHIV-1 RNA v.1.2 kit and expressed as copies per gram of tissueor milliliters of plasma.

Statistical AnalysisAll statistical analyses were performed in Prism 5 from Graph-Pad Software (La Jolla, CA). Paired observations were com-pared using the Wilcoxon matched pairs tests or the Friedmantest with the Dunn posttest. Linear correlations were assessedusing a Spearman rank correlation coefficient. Two-tailed Pvalues of >.05 were considered statistically significant.

RESULTS

Patient CharacteristicsThe demographic and clinical characteristics of our patientpopulation are summarized in Table 1. Symptoms leading toendoscopy were diarrhea in 27 patients (71.1%) and/or unex-plained weight loss in 26 (68.4%). Infectious agents werepresent in the duodenum and jejunum of 6 patients (Crypto-sporidium in 4 patients and Giardia in 1 patient). Active Schis-tosoma infection was identified in the colon of 1 patient withmild acute colitis. Thirty-one patients had no evidence of op-portunistic infections, as assessed by stool cultures and histo-logical analyses. Nonspecific mild-to-moderate enteritis orcolitis was detected in at least 1 site in 92% patients, withsimilar frequencies in all intestinal sites (77%, 70%, 70%, and67% in the duodenum, jejunum, ileum, and colon, respective-ly). Of the 38 HIV-1–positive patients studied, 71% and 66%had samples collected 3 and 6 months after the initiation ofcART, respectively. Ninety-two percent had good immunologi-cal responses (defined as a >3-fold increase in blood CD4+

T-cell counts), and 75% had complete virological responsesafter 6 months of cART (defined as a plasma viral load of <50copies/mL). Only 5 patients continued to have chronic diarrheaat the 6-month time point.

CD4+ T-Cell Depletion Was More Severe in the Small Versusthe Large IntestineCD4+ T-cell percentages and absolute numbers were signifi-cantly reduced in the gastrointestinal tract of HIV-1–positivepatients, compared with controls. Compared with levels of36.8% ± 8.7% (mean ± standard deviation; in the duodenum)and 39.3% ± 10.4% (in the colon) in controls, CD4+ T-cell per-centages in cART-naive HIV-1–positive patients ranged fromvery low levels in the duodenum (3.8% ± 2.8%) and jejunum(2.6% ± 3.2%) to somewhat higher levels in the ileum (5.1% ±5.4%). CD4+ T-cell percentages in the colon (10.3% ± 8.2%) ofHIV-1–positive patients were significantly higher than in theduodenum (P = .002) and jejunum (P = .001) and comparable tothose in blood (8.4% ± 5.7%; P = .547; Figure 1A). Memory cellsmade up the vast majority (>90%) of CD4+ T cells (Figure 1B).Histological analysis revealed a similar, statistically significant de-crease in the absolute number of CD4+ T cells in HIV-1–positive

Table 1. Clinical Characteristics of 38 Subjects Infected WithHuman Immunodeficiency Virus Type 1 (HIV-1)

Characteristic Value

Sex

Male 15Female 23

Age, y

Mean ± SD 36 ± 9Median (range) 33 (23–58)

CD4+ T-cell count, cells/mm3

Mean ± SD 71 ± 59Median (range) 57 (6–224)

CD4+ T cell percentage

Mean ± SD 8.6 ± 5.5Median (range) 7.4 (0.4–22.7)

Plasma HIV-1 RNA load,log copies/mL

Mean ± SD 5.49 ± 5.73Median (range) 5.09 (2.21–6.48)

Diarrhea

Absent 11Present 27

Weight loss

Absent 6<10% 6

>10% 26

Data are number of patients, unless otherwise indicated. All patients hadheterosexually acquired HIV-1 infection and all were infected with HIV-1subtype C.


patients, compared with controls, with the decrease beingmore severe in the duodenum (78 ± 10 vs 8 ± 5 cells/0.80 mm2;P < .001) than the colon (65 ± 10 vs 17 ± 6 cells/0.8 mm2;P < .001; Figure 1C and 1D). No significant correlations weredetected between CD4+ T-cell counts or percentages in the gutas compared to blood. Compared with controls, CD4+ T-celllevels in the duodenum and colon of HIV-1–positive patientswere reduced by 89.5% and 73.7%, respectively, when assessedby flow cytometry analysis and by 83.3% and 69.4%, respectively,when assessed by immunohistochemistry analysis (Figure 1E).

CD8+ T-Cell Levels Were Higher in the Small Versus the LargeIntestineThe percentages of CD8+ T cells in the duodenum (71.1% ±14.4%) and jejunum (77.5% ± 17.3%) of cART-naive HIV-1–positive patients were significantly higher than in the colon(53.8% ± 18.3%; P = .0069 and P = .0006, respectively), with theileum displaying intermediate levels (66.9% ± 7.1%; Figure 2A).CD8+ T-cell percentages in the colon (53.8% ± 18.3%) werecomparable to those in blood (56.1% ± 8.4%). The absolutenumbers of CD8+ T cells were also significantly higher in theduodenum (101 ± 9.4 cells per 0.80 mm2) as compared to thecolon (74 ± 5.7 cells per 0.80 mm2; P = .012; Figure 2B and 2C).Although not statistically significant, the percentages of activat-ed CD38+CD8+ T cells were slightly higher in the duodenum(35.9% ± 21.42%) and jejunum (40.9% ± 23.1%) than in thecolon (30.1% ± 17.75%; P = .10 and P = .09, respectively;Figure 2D). In the colon, but not in the duodenum, jejunum, orileum, the percentage of activated CD38+CD8+ T cells was posi-tively correlated with the amount of tissue HIV-1 RNA(r = 0.50; P = .04).

HIV-1 RNA Levels Were Higher in the Colon Than in theDuodenum and JejunumAll 38 patients (100%) had detectable HIV-1 RNA in the duo-denum, jejunum, ileum, and colon prior to the initiation ofcART (range, 3.89–9.31 log10 HIV-1 RNA copies/g of tissue).Mean HIV-1 RNA levels were significantly higher in the colon(7.2 ± 1.2 log10 HIV-1 RNA copies/g of tissue) than in the duo-denum (6.4 ± 1.0 log10 HIV-1 RNA copies/g of tissue; P =.0067) and jejunum (6.1 ± 1.1 log10 HIV-1 RNA copies/g oftissue; P = .0004), with intermediate levels in the ileum (6.7 ±0.95 log10 HIV-1 RNA copies/g of tissue; P = .1192; Figure 3A).Normalization by absolute CD4+ T-cell count indicated thathigher levels of HIV-1 RNA in the colon were associated withan increased frequency of CD4+ T cells rather than with an in-crease in virus production per cell (Figure 3B).

NRTI/NNRTI Treatment Resulted in Partial Restoration of CD4+

T-Cell Counts in the Blood and Colon but Not in the DuodenumTo assess the efficacy of short-term cART on CD4+ T-cellcount restoration, we compared CD4+ T-cell levels in peripher-al blood and longitudinal biopsy samples (from the duodenum

and colon) collected before and 3 and 6 months after cART ini-tiation. These analyses were limited to 23 patients who hadcomplete virological responses in plasma (defined as an HIV-1RNA load of <50 copies/mL) and no evidence of opportunisticinfection. The percentages of CD4+ T cells in the blood andcolon of cART-treated HIV-1–positive patients increased from8.4% ± 5.7% and 10.3% ± 8.2%, respectively, at baseline to17.8% ± 10.1% (a 2.1-fold increase; P = .01) and 15.8% ± 7.4%(a 1.5-fold increase; P = .0003), respectively, 6 months after theinitiation of cART (Figure 4A). Blood CD4+ T-cell countsincreased, on average, 3.5-fold after 6 months of cART. Des-pite these increases, the frequency of CD4+ T cells in thesecompartments remained lower than in uninfected controls(17.8% ± 10.1% vs 38.0% ± 9.4% in blood [P = .0042] and15.8 ± 7.4 vs 36.3% ± 10.4% in the colon [P = .0006]). CD4+ T-cell depletion was more persistent in the duodenum, with nomeasurable increase detected after 6 months of cART (meanfrequency, 3.7% ± 2.8% at baseline vs 3.9% ± 3.1% at month 6;P = .5623; Figure 4).

Failed Restoration of CD4+ T-Cell Counts Was Associated WithPersistent Inflammation and CD8+ T-Cell ActivationTo investigate whether cART resulted in decreased inflamma-tion and CD8+ T-cell activation, we examined infiltration ofmononuclear cells, including macrophages, lymphocytes, eo-sinophils, and/or plasma cells, and assessed CD38+CD8+ T-celllevels in longitudinal duodenal and colonic biopsy specimenscollected before cART initiation, 3 months after cART initia-tion (from the duodenum), and 6 months after cART initiation(from the duodenum and colon). Before cART initiation, 67%and 77% of patients had mild-to-moderate levels of nonspecificinflammation in the colon and duodenum, respectively. After 6months of cART, only 44% of HIV-1–positive patients had sus-tained nonspecific colitis (in the colon), compared with 72%with nonspecific enteritis (in the duodenum). Similarly, theproportion of activated CD8+ T cells in the colon decreasedfrom 30.1% ± 17.8% at baseline to 18.9% ± 8.7% (P = .048) after6 months of cART. In the colon, the change in HIV-1 RNAload was positively correlated with the decrease in activatedCD38+CD8+ T-cell percentage (r =−0.38; P = .03). Althoughnot statistically significant, CD38+CD8+ T-cell percentages inthe duodenum were somewhat higher after 3 months of cART(48.6% ± 23.4%) and 6 months of cART (44.1% ± 23.8%) thanat baseline (35.9% ± 21.4%; P = .43 and 0.58, respectively). Thisoccurred despite a marked decrease in the HIV-1 RNA load(from 6.4 ± 1.0 log10 copies/g tissue at baseline to 4.2 ± 0.8 and4.9 ± 1.1 log10 copies/g tissue 3 months [P = .0005] and 6months [P = .0134], respectively, after cART initiation). Thisdecrease was comparable to that observed in the colon, wherethe HIV-1 RNA load decreased from 7.2 ± 1.2 log10 copies/g oftissue at baseline to 5.15 ± 1.2 log10 copies/g of tissue at month6 (P = .0005).


DISCUSSION

The impact of cART on CD4+ T-cell count reconstitution in theintestine of HIV-1–positive patients is a topic under considerable

debate. Contributing to the confusion are the interstudy differ-ences in sampling sites, treatment regimens, cART durations,and disease stages, as well as limitations inherent in cross-sectional studies. To our knowledge, this is the first longitudinal

Figure 1. Depletion of CD4+ T-cell counts in combination antiretroviral therapy (cART)–naive patients with late-stage human immunodeficiency virustype 1 (HIV-1) infection was more severe in the small intestine than in the large intestine. A, Proportion of T cells that express CD4, as measured by flowcytometric analysis of peripheral blood mononuclear cells or total cell suspensions isolated from the gastrointestinal tract of cART-naive patients withAIDS. Mean values and statistical significance among the different sites are indicated (*P < .05, **P < .01, ***P < .001). B, The proportions of naive andmemory CD4+ T cells in the various intestinal sites in a subset of 12 patients prior to the initiation of cART. Memory CD4+ T cells were defined as cellsthat express CD45RO, and naive cells were defined as CD45RO– and CD27+. C, Representative immunohistochemical staining for CD4 (brown) in the duode-num and colon of a cART-naive AIDS patient. D, Decrease in the absolute number of CD4+ T cells in the duodenum and colon of cART-naive patients withAIDS, relative to uninfected controls, as assessed by immunohistochemistry analysis (**P < .01). E, Comparison of the decrease in CD4+ T-cell counts incART-naive patients with AIDS versus uninfected controls as assessed by 2 different methods—flow cytometry (percentage of CD4+ T cells) and histologi-cal analysis (absolute number of CD4+ T cells).


study to examine the extent of CD4+ T-cell count depletion andrestoration in the gastrointestinal tract of African patients withlate-stage disease who had unexplained chronic diarrhea and/orweight loss. In the developing world, individuals presenting withCD4+ T-cell counts of <200 cells/mm3 are fast-tracked for treat-ment with a standardized NRTI-NNRTI regimen [21, 22].

Before cART initiation, we detected a profound decrease inCD4+ T cell counts and a corresponding moderate increase inCD8+ T cell counts in all sites of the gastrointestinal tract, re-gardless of the method (ie, flow cytometry or immunohisto-chemistry analysis) used to quantify T-cell subsets. The severityof depletion exceeded that reported for patients with acute/

early HIV-1 infection, presumably a reflection of the continuedloss of CD4+ T cells during disease progression [14, 23]. One ofthe most striking findings was the profound depletion of CD4+

T cells in the small intestine (ie, duodenum, jejunum, andileum) versus the large intestine (ie, colon) and peripheralblood. Other studies have reported that CD4+ T-cell count de-pletion is more severe in the duodenum [23], ileum [1, 24], andcolon [4, 24] than in the blood of treatment-naive patients andin the duodenum, compared with the colon and rectum [25], ofpatients receiving suppressive cART.

To determine whether the distinct immune environmentssupport different levels of HIV-1, we compared total HIV-1

Figure 2. Increased frequency of CD8+ T cells and activated CD38+CD8+ T cells in the upper small intestine as compared to the colon of combination an-tiretroviral therapy (cART)–naive patients with AIDS. A, Proportion of T cells that express CD8, as measured by flow cytometric analysis of peripheral bloodmononuclear cells or total cell suspensions isolated from the gastrointestinal tract of treatment-naive patients with AIDS. Mean values and statistical sig-nificance among the different sites are indicated (*P < .05, ***P < .001). B, Increase in the absolute number of CD8+ T cells in the duodenum and colon oftreatment-naive patients with AIDS compared to uninfected controls, as quantified by immunohistochemistry. C, Representative immunohistochemicalstaining for CD8 (brown) in the duodenum and colon of a treatment-naive patient with AIDS. D, Proportion of CD8+ T cells expressing the activation markerCD38, as assessed by flow cytometry of cell suspensions isolated from different regions of the gastrointestinal tract of cART-naive patients with AIDS.


RNA levels (encompassing both cell-associated and virionRNA) in the duodenum, jejunum, and ileum with those in thecolon. HIV-1 RNA was readily detectable throughout the gas-trointestinal tract of cART-naive patients, at levels exceedingthose in blood (on the assumption that 1.0 g of tissue is equiva-lent to 1.0 mL of plasma). As observed for CD4+ T-cell counts,HIV-1 RNA levels were moderate in the duodenum andjejunum, intermediate in the ileum, and significantly higher inthe colon. Similarly, in cART-treated patients, studies haveshown that the amount of HIV-1 proviral DNA per CD4+ Tcell increases when descending through the gastrointestinaltract toward the colon and rectum [25].

The positive correlation observed between activatedCD38+CD8+ T-cell counts and HIV-1 RNA loads in the colonof treatment-naive patients, as well as the parallel decrease inboth parameters in response to cART, suggest that CD8+ T cellactivation is closely linked to viral replication. Other investigators

have reported that, in cART-treated patients with maximallysuppressed viral replication, the degree of CD8+ T-cell activationin the colon correlates with the size of the residual HIV-1 DNAreservoir [2]. In the duodenum, the lack of a correlation betweenCD38+CD8+ T-cell counts and HIV-1 RNA loads and the per-sistence of activated CD38+CD8+ T cells during cART, despite amarked decrease in the HIV-1 RNA load, suggest that factorsother than or in addition to HIV-1 infection (eg, coinfections,microbial translocation, disruption of the microenvironment,and persistent immune activation that is not fully reversed bycART) contribute to CD8+ T-cell activation in this tissue.

In agreement with a growing body of evidence [14, 26–28],our results also suggest that persistent inflammation and CD8+

T-cell activation are associated with lower increases in CD4+

T-cell counts. This view is supported by studies showing aninverse correlation between activated CD38+CD8+ T cells andCD4+ T-cell levels in blood [29, 30] and between the activatedKi67+CD8+ T cells and the severity of CD4+ T-cell depletion inthe lower gastrointestinal tract [24]. Although the mechanism(s)are not fully understood, activated CD8+ T cells have the capacityto delay or prevent CD4+ T-cell recovery by producing proin-flammatory cytokines and chemokines, suppressing CD4+ T-cellproliferation, and inducing CD4+ T-cell apoptosis [18, 31, 32].Chronic inflammation and immune activation may also lead toincreased collagen deposition and fibrotic damage [33, 34]. Fi-brotic deposition of collagen occurs in Peyer’s patches of theileum, and the extent of deposition/damage is predictive ofCD4+ T-cell count restoration [35, 36]. To our knowledge, thereare no studies of collagen deposition in the duodenum,jejunum, or colon. Other investigators have reported that genesinvolved in inflammation and cell activation are upregulated inthe jejunum of patients who exhibit poor mucosal CD4+ T-cellresponses [28]. Further studies are needed to determine whetherthese changes are confined to the small intestine.

Another factor that could contribute to poor reconstitutionof the CD4+ T-cell count in the duodenum relates to theCCL25/CCR9 axis of CD4+ T-cell recruitment. This axis playsan important role in lymphocyte homing to the small intestinebut not to the colon. CCL25 is a chemokine produced by epi-thelial cells of the small intestine [37, 38]. Its receptor, CCR9, isexpressed on lymphocytes that traffic to the small intestine [39,40]. HIV-1–positive patients produce less CCL25, and as aresult many CCR9+CD4+ T cells remain in the circulation ratherthan repopulating effector sites in the small intestine [41]. Re-duced homing of CCR9+CD4+ T cells has been associated withmucosal damage, increased microbial translocation, and sys-temic immune activation, changes that may contribute to poorimmune restoration [42].

The strengths of our study, in comparison to previousstudies, relate to the homogeneity of our patient cohort (withrespect to treatment, stage of infection, and, in the majority ofpatients, the absence of overt opportunistic infections) and the

Figure 3. Before the initiation of combination antiretroviral therapy(cART), human immunodeficiency virus type 1 (HIV-1) RNA levels werehigher in the colon than in the upper small intestine of African patientswith AIDS. A, HIV-1 RNA copy numbers in cART-naive patients with AIDS(n = 38) were measured in plasma and tissue extracts prepared from differ-ent sites in the gastrointestinal tract. Biopsy samples collected from differ-ent sites in the gastrointestinal tract were minced with a razor blade,extracted overnight in guanidinium lysis buffer, and processed for HIV-1RNA quantification using the Nuclisens Easy Q assay. Results were report-ed as log10 HIV-1 RNA copies per gram of tissue or per milliliter ofplasma. B, HIV-1 RNA levels in the duodenum and colon, normalized bytheir respective absolute CD4+ T-cell counts.


repeat sampling of multiple tissue sites from the same patientsbefore and after the initiation of cART. We focused on theeffects of short-term cART because we were interested in iden-tifying early predictors of immunological failure. Althoughthere is no consensus as to the best time to assess immunologi-cal responses [43, 44], studies performed on peripheral bloodafter 3–6 months of cART have been shown to be predictive ofimmune restoration in the long-term [45]. Limitations of ourstudy relate to the small size of our control group and thelimited number of immune and virological correlates exam-ined. Furthermore, despite efforts to exclude patients withenteric coinfections (stool cultures and histological staining foracid-fast bacilli and fungal and parasitic infections), we cannot

exclude the possibility that other coinfections may have affectedour results. In addition, cell numbers did not allow us to deter-mine the antigen specificity and clonal diversity of our CD8+ Tcells. Further studies are also required to determine whetherour results can be generalized to patients without diarrhea andweight loss and to patients of non-African origin.

Our study demonstrates that, despite the late stage of infection,gradual restoration of CD4+ T-cell levels is possible in the colonand blood of African patients with AIDS during short-termcART. It also underscores the distinct differences in the smallversus large intestine with regard to CD4+ T-cell depletionand reconstitution and highlights the need for detailed compar-ative studies to better delineate the mechanisms leading to

Figure 4. Impaired CD4+ T-cell restoration and sustained CD8+ T-cell activation in the duodenum, but not in the colon, of patients with AIDS who weretreated with nucleoside reverse-transcriptase inhibitors or nonnucleoside reverse-transcriptase inhibitors. A, Spaghetti plots of the kinetics of CD4+ T-cellrestoration, HIV RNA clearance and CD8+ T-cell activation in the duodenum, colon, and blood. Each line represents an individual patient. Analysis waslimited to patients with a complete virological response at month 6 after cART initiation (plasma viral load, < 50 copies/mL; n = 23). B, Summary statisticsof the kinetics of CD4+ T-cell restoration, human immunodeficiency virus type 1 (HIV-1) RNA clearance, and CD8+ T-cell activation in the duodenum, colon,and blood (n = 23). Results are reported as mean and SD.


immunological failure in the upper small intestine. An under-standing of these mechanisms may lead to the development ofhighly targeted treatment strategies that enhance recovery ofthe CD4+ T-cell count. If inflammation and immune activationin the small intestine are driven by local factors not directlyrelated to HIV-1 (eg, tissue damage, loss of epithelial integrity,and nonspecific activation of CD8+ T cells), it is unlikely thatdrugs targeting HIV replication will be maximally effectivewith respect to immune recovery. This view is supported bystudies showing that treatment intensification with raltegravir,a potent integrase inhibitor, did not have any significant effecton CD4+ T-cell levels or immune activation in the gastrointesti-nal tract or on HIV-1 RNA levels in plasma [46–48]. cART pro-tocols that include immune modulating and antiinflammatoryagents may offer an attractive alternative approach.

Notes

Acknowledgments. We thank the endoscopy nurses of the Unitas andPretoria East GI units and Dr Mark Theron, the study anesthetist for theirdedication and commitment to patient care; and our patients and volun-teers, without whose participation and commitment this study would nothave been possible.E. C. and T. R. designed the study protocol. E. C. conceived and performed

flow cytometric analyses, interpreted the data, and authored the first draft ofthe manuscript. T. R. and S. v. d. M. obtained specimens, supervised the col-lating of clinical data, and cared for the patients. S. M. performed some flowcytometric analyses. T. S. performed all histological analyses, including quan-tification of CD4+ and CD8+ T-cell levels in LM tissue sections. E. C., S. M.,and P. M. performed viral load analyses. G. P. provided immunological ex-pertise and participated in discussions of the results. C. S. provided expertisein database management. S. C. and C. S. obtained funding.Financial support. This work was supported by the Delegation of the

European Union to South Africa (grant Sante 2007/147-790, Drug Resis-tance Surveillence and Treatment Monitoring Network for Public SectorHIV Antiretroviral Treatment Programme in the Free State); the NationalResearch Council of South Africa (Unlocking the Future grant 61509).Potential conflicts of interest. All authors: No reported conflicts.All authors have submitted the ICMJE Form for Disclosure of Potential

Conflicts of Interest. Conflicts that the editors consider relevant to thecontent of the manuscript have been disclosed.

References

1. Brenchley JM, Schacker TW, Ruff LE, et al. CD4+ T cell depletionduring all stages of HIV disease occurs predominantly in the gastroin-testinal tract. J Exp Med 2004; 200:749–59.

2. Fukazawa Y, Miyake A, Ibuki K, et al. Small intestine CD4+ T cells areprofoundly depleted during acute simian-human immunodeficiencyvirus infection, regardless of viral pathogenicity. J Virol 2008; 82:6039–44.

3. Mehandru SM, Poles MA, Tenner-Racz K, et al. Primary HIV-1 infec-tion is associated with preferential depletion of CD4+ T lymphocytesfrom effector sites in the gastrointestinal tract. J Exp Med 2004; 200:761–70.

4. Mehandru S, Poles MA, Tenner-Racz K, et al. Mechanisms of gastroin-testinal CD4+ T-cell depletion during acute and early human immuno-deficiency virus type 1 infection. J Virol 2007; 81:599–612.

5. Li Q, Duan L, Estes JD, et al. Peak SIV replication in resting memoryCD4+ T cells depletes gut lamina propria CD4+ T cells. Nature 2005;434:1148–52.

6. Brenchley JM, Price DA, Schacker TW, et al. Microbial translocation isa cause of systemic immune activation in chronic HIV infection. NatMed 2006; 12:1365–71.

7. Paiardini M, Frank I, Pandrea I, et al. Mucosal immune dysfunction inAIDS pathogenesis. AIDS Rev 2008; 10:36–46.

8. Brenchley JM, Douek DC. HIV infection and the gastrointestinalimmune system. Mucosal Immunol 2008; 1:23–30.

9. Gordon SN, Klatt NR, Bosinger SE, et al. Severe depletion of mucosalCD4+ T cells in AIDS-free simian immunodeficiency virus-infectedsooty mangabeys. J Immunol 2007; 179:3026–34.

10. Pandrea IV, Gautam R, Ribeiro M, et al. Acute loss of intestinal CD4+

T cells is not predictive of simian immunodeficiency virus virulence.J Immunol 2007; 179:3035–46.

11. Brenchley JM, Paiardini M, Knox KS, et al. Differential Th17 CD4T-cell depletion in pathogenic and nonpathogenic lentiviral infections.Blood 2008; 112:2826–35.

12. Aujla SJ, Dubin PI, Kolls JK. Th17 cells and mucosal host defense.Semin Immunol 2007; 19:377–82.

13. Liang SC, Tan XY, Luxenberg DP, et al. Interleukin (IL)-22 and IL-17are coexpressed by Th17 cells cooperatively enhance expression of anti-microbial peptides. J Exp Med 2006; 203:2271–9.

14. Mehandru S, Poles K, Tenner-Racz P, et al. Lack of mucosal immunereconstitution during prolonged treatment of acute and early HIV-1 in-fection. PLoS Med 2006; 3:e484.

15. Gazzola L, Tincati C, Bellistre GM, et al. The absence of CD4+ T cellcount recovery despite receipt of virologically suppressive highly activeantiretroviral therapy: clinical risk, immunological gaps, and therapeu-tic options. Clin Infect Dis 2009; 48:328–37.

16. Tincati C, Biasin M, Bandera A, et al. Early initiation of highly activeantiretroviral therapy fails to reverse immunovirological abnormalitiesin gut-associated lymphoid tissue induced by acute HIV infection.Antivir Ther 2009; 14:321–30.

17. Talal AH, Monard S, Vesanen M, et al. Virologic and immunologiceffect of antiretroviral therapy on HIV-1 in gut-associated lymphoidtissue. J Acquir Immune Defic Syndr 2001; 26:1–7.

18. Guadalupe M, Reay E, Sankaran S, et al. Severe CD4+ T-cell depletionin gut lymphoid tissue during primary human immunodeficiency virustype 1 infection and substantial delay in restoration following highlyactive antiretroviral therapy. J Virol 2003; 77:11708–17.

19. Ciccone EJ, Read SW, Mannon PJ, et al. Cyclcing of gut mucosal CD4+

T cells decreases after prolonged anti-retroviral therapy and is associat-ed with plasma LPS levels. Mucosal Immunol 2010; 3:172–81.

20. Sheth PM, Chege D, Shin LY, et al. Immune reconstitution in thesigmoid colon after long-term HIV therapy. Mucosal Immunol 2008;1:382–8.

21. National Department of Health Republic of South Africa. Clinicalguidelines for the management of HIV & AIDS in adults and adoles-cents, 2004. http://www.sanac.org.za. Accessed May 2012.

22. WHO/UNICEF/UNAIDS. Joint reporting tool on the health sector re-sponse to HIV/AIDS 2010. Geneva, Switzerland: WHO, 2010.

23. Schneider T, Jahn H-U, Schmidt W, et al. Loss of CD4 T lymphocytesin patients infected with human immunodeficiency virus type 1 is morepronounced in the duodenal mucosa than in the peripheral blood. Gut1995; 37:524–9.

24. Gordon SN, Cervasi B, Odorizzi P, et al. Disruption of intestinal CD4+

T cell homeostasis is a key marker of CD4+ T cell activation in HIV-in-fected individuals. J Immunol 2010; 185:5169–79.

25. Yuki SA, Gianella S, Sinclair E, et al. Differences in HIV burden andimmune activation within the gut of HIV-positive patients receivingsuppressive antiretroviral therapy. J Infect Dis 2010; 202:1553–61.

26. Hunt PW, Martin JN, Sinclair E, et al. T cell activation is associatedwith lower CD4+ T cell gains in human immunodeficiency virus-infectedpatients with sustained viral suppression during antiretroviral therapy.J Infect Dis 2003; 187:1534–43.

27. Mavigner M, Delobel P, Cazabat M, et al. HIV-1 residual viremia corre-lates with persistent T-cell activation in poor immunological respond-ers to combination antiretroviral therapy. PLoS One 2009; 4:e7658.


http://www.sanac.org.za



28. Guadalupe M, Sankaran S, George MD, et al. Viral suppression andimmune restoration in the gastrointestinal mucosa of human immuno-deficiency virus type 1-infected patients initiating therapy duringprimary or chronic infection. J Virol 2006; 80:8236–47.

29. Giorgi JV, Liu Z, Hultin LE, et al. Elevated levels of CD38+CD8+ T cellsin HIV infection add to the prognostic value of low CD4+ T cell levels:Results of 6 years of follow-up. The Los Angeles Center, MulticenterAIDS Cohort Study. J Acquir Immune Defic Syndr 1993; 6:904–12.

30. Chun TW, Justement JS, Sanford C, et al. Relationship between the fre-quency of HIV-specific CD8+ T cells and the level of CD38+CD8+ Tcells in untreated HIV-infected individuals. Proc Natl Acad Sci U S A2004; 101:2462–9.

31. Nobel A, Pestano GA, Cantor H. Suppression of immune responses byCD8 cells. I. Superantigen-activated CD8 cells induce unidirectionalFas-mediated apoptosis of antigen-activated CD4 cells. J Immunol1998; 160:559–65.

32. Smit-McBride Z, Mattapallil JJ, Villinger F, et al. Intracellular cytokineexpression in the CD4+ and CD8+ T cells from intestinal mucosa ofsimian immunodeficiency virus infected macaques. J Med Primatol1998; 27:129–40.

33. Schacker TW, Brenchley JM, Beilman GJ, et al. Lymphatic tissue fibro-sis is associated with reduced numbers of naïve CD4+ T cells in humanimmunodeficiency virus type 1 infection. Clin Vaccine Immunol 2006;13:556–60.

34. Estes JD, Haase AT, Schacker TW. The role of collagen deposition indepleting CD4+ T cells and limiting reconstitution in HIV-1 and SIVinfections through damage to secondary lymphoid organ niche. SeminImmunol 2008; 20:181–6.

35. Schacker TW, Cavan R, Beilman GJ, et al. Amount of lymphatic tissuefibrosis in HIV infection predicts magnitude of HAART-associatedchange in peripheral CD cell count. AIDS 2005; 19:2169–71.

36. Estes JD, Baker JV, Brenchley JM, et al. Collagen deposition limitsimmune reconstitution in the gut. J Infect Dis 2008; 198:456–64.

37. Kunkel EJ, Campbell JJ, Haraldsen G, et al. Lymphocyte CC chemokinereceptor 9 and epithelial-thymus expressed chemokine (TECK) expres-sion distinguish the small intestinal compartment: Epithelial expressionof tissue-specific chemokines as an organizing principle in regional im-munity. J Exp Med 2000; 192:761–8.

38. Papadakis KA, Prehn J, Moreno ST, et al. CCR9-positive lymphocytesand thymus-expressed chemokine distinguish small bowel from colonicCrohn’s disease. Gastroenterology 2001; 121:246–54.

39. Johansson-Lindbom B, Svensson M, Wurbel MA, et al. Selective gener-ation of gut tropic T cells in gut-associated lymphoid tissue (GALT): re-quirement for GALT dendritic cells and adjuvant. J Exp Med 2003;198:963–9.

40. Mora JR, Bono MR, Manjunath N, et al. Selective imprinting of gut-homing T-cells by Peyer’s patch dendritic cells. Nature 2003; 424:88–93.

41. Mavigner M, Cazabat M, Dubois M, et al. Altered CD4+ T cell homingto the gut impairs mucosal immune reconstitution in treated HIV-infected individuals. J Clin Invest 2012; 122:62–9.

42. Dronda R, Moreno S, Moreno A, et al. Long-term outcomes among an-tiretroviral-naive human immunodeficiency virus-infected patientswith small increases in CD4+ cell counts after successful virologic sup-pression. Clin Infect Dis 2002; 35:1005–9.

43. Yeni PG, Hammer SM, Hirsch MS, et al. Treatment for adult HIV in-fection: 2004 recommendations of the International AIDS Society-USAPanel. JAMA 2004; 292:251–65.

44. Kaufmann GR, Furrer H, Ledergerber B, et al. Characteristics, determi-nants and clinical relevance of CD4 T cell recovery to <500 cells/μL inHIV type 1- infected individuals receiving potent antiretroviral therapy.Clin Infect Dis 2005; 41:361–72.

45. D’Arminio Monforte A, Testori V, Adorni F, et al. CD4 cell counts atthe third month of HAART may predict clinical failure. AIDS 1999;13:1669–76.

46. Asmuth DM, Ma Z-M, Mann S, et al. Gastrointestinal-associated lym-phoid tissue immune reconstitution in a randomized clinical trial ofraltegravir versus non-nucleoside reverse transcriptase inhibitor-basedregimens. AIDS 2012; 26:1625–34.

47. Hatano HH, Hayes TL, Dahl V, et al. A randomized, controlled trial ofraltegravir intensification in antiretroviral-treated, HIV-infected patientswith a suboptimal CD4+ T cell response. J Infect Dis 2011; 203:960–8.

48. Yukl SA, Shergill AK, McQuaid K, et al. Effect of raltegravir-containingintensification on HIV burden and T-cell activation in multiple gutsites of HIV-positive adults on suppressive antiretroviral therapy. AIDS2010; 24:2451–60.


Impaired CD4+ T cell restoration in the small vs. large intestine of HIV-1+ South Africans receiving...

Documents

Transcript of Impaired CD4+ T cell restoration in the small vs. large intestine of HIV-1+ South Africans receiving...